CloudWatch alarms in AWS - Commands & Configuration

Introduction

CloudWatch alarms watch your cloud resources and tell you when something needs attention. They help you catch problems early by sending alerts based on rules you set.

When you want to know if your server CPU is too high for a long time

When you need to get notified if your website is not responding

When you want to track if your database storage is almost full

When you want to automatically fix problems by triggering actions when an alarm fires

When you want to keep an eye on costs by monitoring resource usage

Config File - cloudwatch-alarm.json

cloudwatch-alarm.json

{
  "AlarmName": "HighCPUUsageAlarm",
  "AlarmDescription": "Alarm when CPU exceeds 80% for 5 minutes",
  "ActionsEnabled": true,
  "AlarmActions": ["arn:aws:sns:us-east-1:123456789012:NotifyMe"],
  "MetricName": "CPUUtilization",
  "Namespace": "AWS/EC2",
  "Statistic": "Average",
  "Dimensions": [
    {
      "Name": "InstanceId",
      "Value": "i-0abcdef1234567890"
    }
  ],
  "Period": 300,
  "EvaluationPeriods": 1,
  "Threshold": 80.0,
  "ComparisonOperator": "GreaterThanThreshold"
}

This JSON file defines a CloudWatch alarm named HighCPUUsageAlarm. It watches the average CPU usage of a specific EC2 instance over 5 minutes (300 seconds). If the CPU goes above 80%, the alarm triggers. It sends a notification to an SNS topic NotifyMe. The Dimensions specify which EC2 instance to watch by its ID.

Commands

This command creates or updates the CloudWatch alarm using the JSON file. It sets the rules and notification actions.

Terminal

aws cloudwatch put-metric-alarm --cli-input-json file://cloudwatch-alarm.json

Expected Output

No output (command runs silently)

→

--cli-input-json - Specifies the JSON file with alarm configuration

This command checks the details and status of the alarm named HighCPUUsageAlarm to confirm it was created.

Terminal

aws cloudwatch describe-alarms --alarm-names HighCPUUsageAlarm

Expected Output

{

    "MetricAlarms": [

        {

            "AlarmName": "HighCPUUsageAlarm",

            "AlarmDescription": "Alarm when CPU exceeds 80% for 5 minutes",

            "ActionsEnabled": true,

            "OKActions": [],

            "AlarmActions": [

                "arn:aws:sns:us-east-1:123456789012:NotifyMe"

            ],

            "InsufficientDataActions": [],

            "StateValue": "OK",

            "StateReason": "Threshold Crossed: 1 datapoint [0.0 (27/04/24 10:00:00)] was not greater than the threshold (80.0).",

            "MetricName": "CPUUtilization",

            "Namespace": "AWS/EC2",

            "Statistic": "Average",

            "Dimensions": [

                {

                    "Name": "InstanceId",

                    "Value": "i-0abcdef1234567890"

                }

            ],

            "Period": 300,

            "EvaluationPeriods": 1,

            "Threshold": 80.0,

            "ComparisonOperator": "GreaterThanThreshold"

        }

    ]

}

→

--alarm-names - Filters output to show only the specified alarm

This command lists all CPU usage metrics for EC2 instances, helping you find the right instance ID to monitor.

Terminal

aws cloudwatch list-metrics --namespace AWS/EC2 --metric-name CPUUtilization

Expected Output

{

    "Metrics": [

        {

            "Namespace": "AWS/EC2",

            "MetricName": "CPUUtilization",

            "Dimensions": [

                {

                    "Name": "InstanceId",

                    "Value": "i-0abcdef1234567890"

                }

            ]

        }

    ]

}

→

--namespace - Filters metrics by service namespace

→

--metric-name - Filters metrics by metric name

Key Concept

If you remember nothing else from this pattern, remember: CloudWatch alarms watch specific metrics and alert you when those metrics cross thresholds you set.

Common Mistakes

Using the wrong instance ID in the alarm dimensions

The alarm will watch the wrong resource and never trigger for your intended server

Always verify the instance ID by listing metrics or checking your EC2 instances before creating the alarm

Not enabling actions in the alarm configuration

The alarm will detect problems but will not send notifications or trigger actions

Set ActionsEnabled to true and specify AlarmActions like SNS topics to get alerts

Setting the threshold too high or too low without testing

You may get too many false alarms or miss real problems

Start with reasonable thresholds and adjust based on observed metric behavior

Summary

Create a CloudWatch alarm by defining a JSON file with metric, threshold, and actions.

Use the AWS CLI to apply the alarm configuration with put-metric-alarm.

Check alarm status and details with describe-alarms to confirm it works.